The Semantic Taxonomy Induction Project Page

Overview

This project focuses on the automatic acquisition of fundamental semantic relationships between words from corpora. In our most recent work we generalize upon methods presented by (Caraballo, 2001) and (Hearst, 1992) for discovering noun pairs related via hypernymy (the "is-a" relation). We have demonstrated the ability to learn new syntactic patterns for the reliable identification of hypernym noun pairs, and shown significant improvements in hypernym classification over past automatic methods and hand-constructed taxonomies. We are currently working to leverage our capability for high-precision relationship discovery to extend widely-used lexical resources (e.g. WordNet), as well as to automatically construct novel semantic taxonomies directly from corpora.

People

Professors:
- Dan Jurafsky
- Andrew Ng
Students:
- Rion Snow

Papers

Rion Snow, Daniel Jurafsky, and Andrew Y. Ng, "Semantic taxonomy induction from heterogenous evidence". COLING/ACL, 2006. Received Best Paper Award. [pdf]
Rion Snow, Daniel Jurafsky, and Andrew Y. Ng, "Learning syntactic patterns for automatic hypernym discovery". NIPS 17, 2005. [pdf]

Related Work

Sharon Caraballo, "Automatic Acquisition of a Hypernym-Labeled Noun Hierarchy from Text". Brown University Ph.D. Thesis, 2001. [ps]
Marti Hearst, "Automatic Acquisition of Hyponyms from Large Text Corpora". COLING 1992. [pdf].
George Miller, "WordNet: a lexical database for English. Communications of the ACM, 1995. [pdf] [WordNet homepage]

Contact Information

For any comments or questions, please feel free to email Rion.